In the following we discuss observed and simulated column-averaged CH4 mixing ratios over Europe,
North America, and Australia in some more detail (further extending the discussion in sections 4.1
and 4.2).

Europe
Compared to tropical regions, column averaged CH4 mixing ratios over Europe show much smaller
enhancements in the 3-month averages (Figure 5 (scenario S1) and Figure S1c (scenario S3)).
This is due to relatively uniformly distributed sources and the change of wind directions related
to changes of synoptic situations, which is smoothing source plumes over large areas 
during the averaging period. Model simulations, however, show a small gradient from Southwest
to East, with lowest mixing ratios over the Iberian peninsula. This gradient is also visible in the
SCIAMACHY observations. In general, SCIAMACHY data show a somewhat larger scatter than the model
simulations. Despite the overall relatively small average gradient, large gradients can build up
during specific synoptic situations [de Beek et al., 2006]. Therefore, over Europe it will be useful
to analyze composites of shorter time periods (i.e. on the order of days), rather than monthly or
longer averages, to get additional information on the emissions from satellite. The exploitation of
information related to synoptic scale variations has been demonstrated to be very useful also for
quasi-continuous surface measurements [Bergamaschi et al., 2005]. A more detailed investigation of
such synoptic variations in column averaged mixing ratios, however, is beyond the scope of this paper.

North America
Similar to Europe, relatively small enhancements in CH4 mixing ratios are observed in the 3-month
averages over the United States and Canada (Figure 5 (scenario S1) and Figure S1a (scenario S3)).
An exception is very high CH4 values over the north eastern United States and southern Canada during
January to March, which is not reproduced by the model simulations (scenario S1). However, these
high values originate from very few observations only (confined to March, while almost no data are
available for January and February at these high latitudes; furthermore the scatter to these data
is relatively large) and have therefore large uncertainty. In scenario S3 these high observations
are corrected significantly downwards by the applied bias correction. Apart from this discrepancy
the agreement is favorable; in particular, the East-West gradient is relatively consistent between
observations and model simulations. To a large extent this gradient is due to the topography, with
smaller column averaged CH4 mixing ratios over the Rocky Mountains than over lower terrain. During
July to September, some enhancement is visible over the highly industrialized regions of the US East
coast (around 40 degrees N), apparent both in the SCIAMACHY data and the model simulations (but more
prominent in the SCIAMACHY data). In general, however, such enhancements are much more pronounced
over shorter averaging time periods (similar to Europe).

Australia
The spatial gradients over Australia are generally very small throughout the year as a result of
small CH4 emissions. This very small gradient is consistent between SCIAMACHY observations and model
simulations (Figure 5 (scenario S1) and Figure S1f (scenario S3)). However, the uncorrected
SCIAMACHY data show a seasonal cycle of about 30 ppb amplitude, while model simulations are almost
constant throughout the year (scenario S1). Lowest SCIAMACHY data are retrieved during April to June,
concomitant with lower observational data (compared to model simulations) over Africa, South America,
and the ocean at about the same latitude range (~20 to 40 degrees S). This discrepancy has been
noted already by Frankenberg et al. [2006]. This study confirms that this discrepancy persists with
the optimized model simulations, ensuring very good agreement of model simulations and surface
measurements within this latitude region, e.g., at Samoa (SMO), Easter Island (EIC), Crozet Island
(CRZ), and Cape Grim (CGO). In section 4.5 we showed that potential discrepancies between simulated
and real stratospheric CH4 mixing cannot explain the SCIAMACHY-TM5 discrepancies in total columns.
In scenario S3 this seasonal bias is compensated by the applied bias correction, resulting in a very
good agreement of corrected observations and model simulations throughout the year (Figure S1f).

References
Bergamaschi, P., M. Krol, F. Dentener, A. Vermeulen, F. Meinhardt, R. Graul, M. Ramonet, W. Peters,
and E.J. Dlugokencky, Inverse modelling of national and European CH4 emissions using the atmospheric
zoom model TM5, Atmos. Chem. Phys., 5, 2431-2460, 2005.

de Beek, R., M. Buchwitz, S. Noel, J.P. Burrows, H. Bovensmann, M. Bruns, H. Bremer, and P.
Bergamaschi, Atmospheric carbon gases retrieved from SCIAMACHY by WFM-DOAS: Improved global CO
and CH4 and initial verification of CO2 over Park Falls (46oN, 90oW), Atmos. Chem. Phys. Discuss.,
6, 363-399, 2006.

Frankenberg, C., J.F. Meirink, P. Bergamaschi, A.P.H. Goede, M. Heimann, S. Koerner, U. Platt, M.
van Weele, and T. Wagner, Satellite chartography of atmospheric methane from SCIAMACHY on board
ENVISAT: Analysis of the years 2003 and 2004, J. Geophys. Res., 111, D07303,
doi:10.1029/2005JD006235, 2006.